Numerical Solution of partial differential equations

Similar documents
Preface. 2 Linear Equations and Eigenvalue Problem 22

NUMERICAL METHODS FOR ENGINEERING APPLICATION

Index. higher order methods, 52 nonlinear, 36 with variable coefficients, 34 Burgers equation, 234 BVP, see boundary value problems

NUMERICAL COMPUTATION IN SCIENCE AND ENGINEERING

1 Number Systems and Errors 1

Applied Numerical Analysis

Introduction to Applied Linear Algebra with MATLAB

ME Computational Fluid Mechanics Lecture 5

Numerical Methods for Engineers and Scientists

Numerical Methods. Scientists. Engineers

Today s class. Linear Algebraic Equations LU Decomposition. Numerical Methods, Fall 2011 Lecture 8. Prof. Jinbo Bi CSE, UConn

Introduction to Numerical Analysis

Department of Mathematics California State University, Los Angeles Master s Degree Comprehensive Examination in. NUMERICAL ANALYSIS Spring 2015

Numerical Analysis. A Comprehensive Introduction. H. R. Schwarz University of Zürich Switzerland. with a contribution by

Numerical Mathematics

APPLIED NUMERICAL LINEAR ALGEBRA

Applied Linear Algebra

CAM Ph.D. Qualifying Exam in Numerical Analysis CONTENTS

NUMERICAL MATHEMATICS AND COMPUTING

ECE539 - Advanced Theory of Semiconductors and Semiconductor Devices. Numerical Methods and Simulation / Umberto Ravaioli

AIMS Exercise Set # 1

Computation Fluid Dynamics

Numerical Methods for Engineers

Chapter 10 Exercises

Preface to the Second Edition. Preface to the First Edition

Introduction to Computational Fluid Dynamics

Partial Differential Equations

A THEORETICAL INTRODUCTION TO NUMERICAL ANALYSIS

Contents. Preface to the Third Edition (2007) Preface to the Second Edition (1992) Preface to the First Edition (1985) License and Legal Information

STABILITY FOR PARABOLIC SOLVERS

6. Iterative Methods for Linear Systems. The stepwise approach to the solution...

Partial differential equations

Math 411 Preliminaries

Numerical Methods in Matrix Computations

Numerical Methods for Engineers. and Scientists. Applications using MATLAB. An Introduction with. Vish- Subramaniam. Third Edition. Amos Gilat.

Math 7824 Spring 2010 Numerical solution of partial differential equations Classroom notes and homework

Chapter 5 HIGH ACCURACY CUBIC SPLINE APPROXIMATION FOR TWO DIMENSIONAL QUASI-LINEAR ELLIPTIC BOUNDARY VALUE PROBLEMS

Finite Difference Method for PDE. Y V S S Sanyasiraju Professor, Department of Mathematics IIT Madras, Chennai 36

Computer Aided Design of Thermal Systems (ME648)

THE NUMERICAL TREATMENT OF DIFFERENTIAL EQUATIONS

Math 411 Preliminaries

2 Multidimensional Hyperbolic Problemss where A(u) =f u (u) B(u) =g u (u): (7.1.1c) Multidimensional nite dierence schemes can be simple extensions of

Tyn Myint-U Lokenath Debnath. Linear Partial Differential Equations for Scientists and Engineers. Fourth Edition. Birkhauser Boston Basel Berlin

Multi-Factor Finite Differences

FDM for parabolic equations

Contents. Preface... xi. Introduction...

Computational Methods

7 Hyperbolic Differential Equations

Math Introduction to Numerical Analysis - Class Notes. Fernando Guevara Vasquez. Version Date: January 17, 2012.

Numerical Solutions to Partial Differential Equations

Numerical methods Revised March 2001

2.29 Numerical Fluid Mechanics Spring 2015 Lecture 9

Chapter Two: Numerical Methods for Elliptic PDEs. 1 Finite Difference Methods for Elliptic PDEs

Finite Difference and Finite Element Methods for Solving Elliptic Partial Differential Equations

x x2 2 + x3 3 x4 3. Use the divided-difference method to find a polynomial of least degree that fits the values shown: (b)

Introduction to Numerical Analysis

Iterative Methods. Splitting Methods

Introduction. Finite and Spectral Element Methods Using MATLAB. Second Edition. C. Pozrikidis. University of Massachusetts Amherst, USA

Numerical Methods with MATLAB

Study of Effect in Convergence of a Solution with the Variation in Accelerating Factor in Successive Over Relaxation Method

ADVANCED ENGINEERING MATHEMATICS

Scientific Computing: An Introductory Survey

Notes on Numerical Analysis

R. Courant and D. Hilbert METHODS OF MATHEMATICAL PHYSICS Volume II Partial Differential Equations by R. Courant

SOLUTION OF EQUATION AND EIGENVALUE PROBLEMS PART A( 2 MARKS)

TABLE OF CONTENTS INTRODUCTION, APPROXIMATION & ERRORS 1. Chapter Introduction to numerical methods 1 Multiple-choice test 7 Problem set 9

Numerical Analysis of Differential Equations Numerical Solution of Parabolic Equations

Numerical Algorithms for Visual Computing II 2010/11 Example Solutions for Assignment 6

Semester I. Mathematics I (Calculus with applications in Chemistry I) Code: MM

10.2 ITERATIVE METHODS FOR SOLVING LINEAR SYSTEMS. The Jacobi Method

FDM for wave equations

ITERATIVE METHODS FOR SPARSE LINEAR SYSTEMS

NUMERICAL METHODS USING MATLAB

KINGS COLLEGE OF ENGINEERING DEPARTMENT OF MATHEMATICS ACADEMIC YEAR / EVEN SEMESTER QUESTION BANK

Matrix Algorithms. Volume II: Eigensystems. G. W. Stewart H1HJ1L. University of Maryland College Park, Maryland

Basic Aspects of Discretization

Numerical Analysis. Elements of. Second Edition. Radhey S. Gupta

NUMERICAL METHODS. lor CHEMICAL ENGINEERS. Using Excel', VBA, and MATLAB* VICTOR J. LAW. CRC Press. Taylor & Francis Group

CN - Numerical Computation

Partial Differential Equations with Numerical Methods

Module 3: BASICS OF CFD. Part A: Finite Difference Methods

APPLIED PARTIAL DIFFERENTIAL EQUATIONS

Numerical Methods - Numerical Linear Algebra

Part 1. The diffusion equation

u n 2 4 u n 36 u n 1, n 1.

Index. C 2 ( ), 447 C k [a,b], 37 C0 ( ), 618 ( ), 447 CD 2 CN 2

LINEAR SYSTEMS (11) Intensive Computation

Numerical Analysis. Introduction to. Rostam K. Saeed Karwan H.F. Jwamer Faraidun K. Hamasalh

A Method for Constructing Diagonally Dominant Preconditioners based on Jacobi Rotations

NUMERICAL SOLUTION OF HYPERBOLIC PARTIAL DIFFERENTIAL EQUATIONS

MATHEMATICS. Course Syllabus. Section A: Linear Algebra. Subject Code: MA. Course Structure. Ordinary Differential Equations

Chapter One: Introduction

A Simple Compact Fourth-Order Poisson Solver on Polar Geometry

Algebra C Numerical Linear Algebra Sample Exam Problems

CLASSICAL ITERATIVE METHODS

Basics of Discretization Methods

Linear Partial Differential Equations for Scientists and Engineers

LINEAR ALGEBRA: NUMERICAL METHODS. Version: August 12,

Numerical Solutions to Partial Differential Equations

Tutorial 2. Introduction to numerical schemes

Transcription:

G. D. SMITH Brunei University Numerical Solution of partial differential equations FINITE DIFFERENCE METHODS THIRD EDITION CLARENDON PRESS OXFORD

Contents NOTATION 1. INTRODUCTION AND FINITE-DIFFERENCE FORMULAE Descriptive treatment of elliptic equations 1 Descriptive treatment of parabolic and hyperbolic equations 4 Finite-difference approximations to derivatives 6 Notation for functions of several variables 8 2. PARABOLIC EQUATIONS: FINITE DIFFERENCE METHODS, CONVERGENCE, AND STABILITY Transformation to non-dimensional form 11 An explicit finite-difference approximation to SU/dt = d 2 U/dx 2 12 A worked example covering three cases and including comparison tables 13 Crank-Nicolson implicit method 19 Worked example including a comparison table 21 Solution of the implicit equations by Gauss's elimination method 24 The stability of the elimination method 27 A weighted average approximation 28 Derivative boundary conditions 29 Worked examples including comparison tables: (i) Explicit formula and central-differenced boundary condition 31 (ii) Explicit formula and forward-differenced boundary condition 33 (iii) Implicit formula and central-differenced boundary condition 36 The local truncation error and a worked example 38 Consistency and a worked example illustrating both consistency and inconsistency 40 Convergence, descriptive treatment and the analysis of an explicit approximation 43 Stability, descriptive treatment 47

Vlll Contents Vector and matrix norms, subordinate matrix norms, p(a)=s l A 49 A necessary and sufficient condition for stability, A =sl, and two worked examples 51 Matrix method of analysis, fixed mesh size. 57 A note on the eigenvalues of /(A) and [/i(a)] _1 /2(A) 58 The eigenvalues of a common tridiagonal matrix 59 Theorems on bounds for eigenvalues and an application. (Gerschgorin's theorems) 60 Gerschgorin's circle theorem and the norm of matrix A 62 Stability criteria for derivative boundary conditions using (i) the circle theorem (ii) A»«1 63 Stability condition allowing exponential growth 66 Stability, von Neumann's method, and three worked examples 67 The global rounding error 71 Lax's equivalence theorem (statement only) and a detailed analysis of a simple case 72 Finite-difference approximations to du/dt = V 2 U in cylindrical and spherical polar co-ordinates 75 A worked example involving lim(3fj/öx)/x 77 Exercises and Solutions 79 PARABOLIC EQUATIONS: ALTERNATIVE DERIVATION OF DIFFERENCE EQUATIONS AND MISCELLANEOUS TOPICS Reduction to a System of ordinary differential equations 111 A note on the Solution of dv/dt = AV + b 113 Finite-difference approximations via the ordinary differential equations 115 The Pade approximants to exp 0 116 Standard finite-difference equations via the Pade approximants 117 A 0 -stability, L 0 -stability and the symbol of the method 119 A necessary constraint on the time step for the Crank- Nicolson method 122 The local truncation errors associated with the Pade approximants 124 Stiff equations 126 An extrapolation method for improving accuracy in t 126 The symbol for the extrapolation method 128 The arithmetic of the extrapolation method 129 The local truncation errors and Symbols of extrapolation schemes 132

Contents The eigenvalue-eigenvector Solution of a System of ordinary differential equations (i) Preliminary results 132 (ii) The eigenvalue-eigenvector Solution of dv/df = AV 134 (iii) An application giving an approximate Solution for large t 135 Miscellaneous methods for improving accuracy: (i) Reduction of the local truncation error - the Douglas equations 137 (ii) Use of three time-level difference equations 138 (iii) Deferred correction method 139 (iv) Richardson's deferred approach to the limit 141 Solution of non-linear parabolic equations: (i) Newton's linearization method and a worked example 142 (ii) Richtmyer's linearization method 144 (iii) Lee's three time-level method 146 A comparison of results for methods (i), (ii), and (iii) for a particular problem 147 The stability of three or more time-level difference equations: (i) A useful theorem on eigenvalues 148 (ii) Matrices with common eigenvector Systems 150 (iii) A worked example 150 Introduction to the analytical Solution of homogeneous difference equations: 153 (i) The eigenvalues and vectors of a common tridiagonal matrix 154 (ii) The analytical Solution of the classical explicit approximation to du/dt = d 2 U/dx 2 156 Exercises and Solutions 158 ix HYPERBOLIC EQUATIONS AND CHARACTERISTICS Analytical Solution of first-order quasi-linear equations 175 A worked example and discussion 176 Numerical Integration along a characteristic 178 A worked example 179 Finite-difference methods on a rectangular mesh for firstorder equations: (i) Lax-Wendroff explicit method and a worked example with a comparison table 181 (ii) Lax-Wendroff method for a set of simultaneous equations 183 (iii) The Courant-Friedrichs-Lewy condition 186

X Contents (iv) Wendroff's implicit approximation 187 Propagation of discontinuities, first-order equations: (i) Discontinuous initial values 188 (ii) Discontinuous initial derivatives 189 Discontinuities and finite-difference approximations. An example using Wendroff's implicit approximation 190 Reduction of a first-order equation to a System of ordinary differential equations 193 The (1,0) Pade difference approximation 195 A comment on the non-stiffness of the equations 196 The (1, 1) Pade or Crank-Nicolson difference equations 196 An improved approximation to du/dx and the (1,0) Pade difference equations 197 A word of caution on the central-difference approximation to du/dx 200 Second-order quasi-linear hyperbolic equations. Characteristic curves, and the differential relationship along them 202 Numerical Solution by the method of characteristics 204 A worked example 207 A characteristic as an initial curve 209 Propagation of discontinuities, second-order equations 210 Finite-difference methods on a rectangular mesh for secondorder equations: 213 (i) Explicit methods and the Courant-Friedrichs-Lewy condition 213 (ii) Implicit methods with particular reference to the waveequation 216 Simultaneous first-order equations and stability 217 Exercises and Solutions 220 5. ELLIPTIC EQUATIONS AND SYSTEMATIC ITERATIVE METHODS Introduction 239 Worked examples: (i) A torsion problem. (ii) A heatconduction problem with derivative boundary conditions 240 Finite-differences in polar co-ordinates 245 Formulae for derivatives near a curved boundary Improvement of the accuracy of Solutions: (i) Finer mesh. (ii) Deferred approach to the limit. (iii) Deferred correction method. (iv) More accurate finite-difference formulae including the nine-point formula 248 Analysis of the discretization error of the five-point approximation to Poisson's equation over a rectangle. Quoted 252 result for irregulär boundaries 254

Contents XI Comments on the Solution of difference equations, covering Gauss elimination, LU decomposition, rounding errors, illconditioning, iterative refinement, iterative methods 257 Systematic iterative methods for large linear Systems 260 Jacobi, Gauss-Seidel, and SOR methods 261 A worked example covering each method 263 Jacobi, Gauss-Seidel, and SOR methods in matrix form 266 A necessary and sufficient condition for convergence of iterative methods 268 A sufficient condition for convergence 269 Asymptotic and average rates of convergence 270 Methods for accelerating convergence. (i) Lyusternik's method. (ii) Aitken's method. An illustrative example 272 Eigenvalues of the Jacobi and SOR iteration matrices and two worked examples 275 The Optimum acceleration parameter for the SOR method. A necessary theorem 277 Proof of (A. + cö 1) 2 = Ao> 2 /j, 2 for block tridiagonal coefficient matrices 279 Non-zero eigenvalues of the Jacobi iteration matrix 280 Theoretical determination of the optimum relaxation parameter w b 282 Calculation of (o b for a rectangle and other Solution domains 285 The Gauss-Seidel iteration matrix H(l) 285 Re-ordering of equations and unknowns 286 Point iterative methods and re-orderings 287 Introduction to 2-cyclic matrices and consistent ordering 288 2-cyclic matrices 289 Ordering vectors for 2-cyclic matrices 290 Consistent ordering of a 2-cyclic matrix 292 The ordering vector for a block tridiagonal matrix 294 An example of a consistently ordered 2-cyclic matrix that is not block tridiagonal 297 Additional comments on consistent ordering and the SOR method 297 Consistent orderings associated with the five-point approximation to Poisson's equation 298 Stone's strongly implicit iterative method 302 A recent direct method 309 Exercises and Solutions 311 INDEX 334